Geographia Technica, Vol. 09, Issue 1, 2014, pp 1 to 9

USING LANDSAT IMAGERY FOR MONITORING THE

SPATIOTEMPORAL EVOLUTION OF SANDING IN DRYLAND, THE

CASE OF IN-SALAH IN THE TIDIKELT (SOUTHERN ALGERIAN

SAHARA)

BOULGHOBRA Nouar1, HADRI Toufik, BOUHANA Mamar

ABSTRACT The exposure to sanding risk in the Tidikelt plain is particularly ascribed to its physiographic and

climatic conditions, the location of agglomerations, agricultural and socioeconomic activities in a

corridor of severe wind action, increases the vulnerability to the risk as a response to the relentless

human pressure, which is exerted on an arid environment already subject to climatic extremes. The use

of remote sensing tool and the multitemporal Landsat imagery have allowed the regional and static

observation of sand accumulation forms, and also monitoring the spatiotemporal dynamics of mobile

dunes during 19 years. This gradual mutation still confronted with a noticeably low resilience from the

population and the authorities.

Key words: Sanding; Mobile dunes; In-Salah; Algeria; Landsat imagery.

1. INTRODUCTION

The international United Nations convention of desertification combating, defines the term

desertification as land degradation in arid, semi-arid and dry sub-humid areas resulting from

various factors, including climatic variations and human activities (UN, 1994). Currently, the

subjected areas to desertification occupy 39.2 % of the globe, i.e. nearly 51.6×106 km² of

degraded land, in which 4.86×106 km² in Africa (Thomas, 1995; Arar, 1993).

Algeria includes large arid and semi arid areas subjected to adverse impacts of desertification;

20×106 ha area is threatened by wind erosion, 5×106 ha area is already in an advanced state

of degradation (Bensaid, 2006). These spaces that are part of the Algerian Sahara which

covering 200×106 ha (80 % of the Algerian territory) have broad aptitudes to the land

degradation, i.e. paleoclimate, drought, long duration of insolation and high wind activity.

This translates to increased wind erosion and therefore, amplifying the sanding risk which

induces not only the decline of ecosystems, but also increased the society vulnerability to

climatic risks.

This is the case of the In-Salah region in southern Algerian Sahara, which offers favorable

physical and climate context to wind erosion. All human and socio-economic installations in

this region are located in a wind corridor of high dynamics, and then present the only stopping

obstacles at the actions of moving sands, hence frequent sanding of agricultural lands, palm

groves, urban agglomerations and road infrastructures.

This article focuses on the diachronic analysis of the sanding evolution in the In-Salah region

using remote sensing tools. The experimental approach involves the study of spatial

extension and bi-temporal evolution of mobile dunes, using multitemporal Landsat imagery;

the TM scene of 1987 and ETM + of 2005. After determining the physiographic and climatic

abilities to sanding in the region, we proceeded to define the various components of the

1 Scientific and technical research center for arid regions (CRSTRA), Touggourt, Algeria,

boulghobra.n@gmail.com, hadtoufik@gmail.com, bouhanamamar@gmail.com

2

landscape (static situation) for each Landsat scene, to finally appreciate the space elements

changing during 19 years, in particular, the mobile dunes (dynamic situation).

2. PHYSIOGRAPHIC AND CLIMATIC CONTEXT OF THE STUDY AREA

Between the longitudes 26o, 28o and latitudes 2o, 3o, In-Salah region is linked administratively

to the Tamanrasset province, and is part of the Tidikelt plain which is a corridor of northeast-

southwest elongation (250×50 km). This topographic depression corresponds to a Mesozoic

sedimentary basin located in the southern piedmont of the Tademaït tray (450 m).

The study area extends over 374 km² and takes position in the Tidikelt (Fig. 1); it includes

the town of In-Salah and neighboring: Sahela Foukania, Sahela Tahtania, Igostène, Hassi El-

Hajar and Foggaret El-Arab. These agglomerations and palm groves sheltering 32,518

inhabitants (NOS, 2008), are parts of a monotonous landscape consisted of vast sand extents,

namely Erg Sidi Moussa, the main source of sands, added to endorheic depressions (Sebkhas)

occupying the lowlands in the region.

Fig. 1 Geolocation of the study area

In-Salah region is characterized by a Saharan climate of hyper-arid type; 39 years series of

climatic data (In-Salah station, 1973-2012), the Demartone aridity index (1925) and the

Emberger bioclimatic quotient (1932) are among the lowest values in Algeria; 0.87 and 0.05

respectively. This aridity is mostly due to the continental character of the Algerian Sahara

marked by the absence of moist air masses revealing a particularly dry area. Precipitation is

scarce and spatiotemporally irregular; the inter-annual normal does not exceed 31.7 mm, with

an intra-annual maximum of 8 mm recorded in April (Fig. 2).

Nouar BOULGHOBRA, Toufik HADRI and Mamar BOUHANA / USING LANDSAT IMAGINARY …3

Fig. 2 Intra-annual variations in temperature, rainfall and humidity, In-Salah station, 1973-2012

With an interannuel mean temperature of 26.4 °C, the In-Salah region is among the hottest

areas in Algeria. The warm season lasts more than 6 months, with a minimum of 5 oC in

January and a maximum often exceeds 45 °C in July. Duration of solar insolation exceeds

3000 hours/year, while procreating low relative humidity value (26.1 %), reaching 40 % only

in January and December.

About the wind dynamics, the central Algerian Sahara (Adrar, In-Salah and Timimoune) is

the windiest region in Algeria (Kasbadji Merzouk, 1999). The average wind speed can reach

5 m/s and exceed 8 m/s for the maximum values, the mean winds class (6-11 m/s) is the most

frequent with 38.5 %. Also, the frequency of strong winds exceeding 11 m/s is 7 % (Fig. 3).

From this threshold, wind becomes efficient and may then perform abrading, transport and

accumulation actions. Directions are multiples, but two dominants directions are

distinguished (Fig. 4): the east winds (Guebli) and northeast (continental trade wind) with

the frequencies 30 % and 22 % respectively (Kerbani, 2013).

Fig. 3 Frequency distribution of wind

classes, In-Salah station, 1998-2008

Fig. 4 Compass rose and annual distribution of wind

directions, In-Salah station 1998-2008 (Kerbani,

2013)

In general, the study area is very favorable to the wind erosion and sanding, since it offers

dry soils, flat and bare lands and favorable climatic context (FAO, 2010). The linear dunes

and barchans (20 meters) of Erg Sidi Moussa are the predominant accumulation forms and

confirm the importance of the wind dynamics occurred in the region. The direction of

4

frequent and efficient wind is in right coincidence with the Tidikelt elongation direction, so

all human and socio-economic installations are located in a high wind dynamics corridor, and

subsequently represent the only obstacles to the shifting sands. This is why the frequent and

constant sanding of palm groves, habitations and infrastructures (Fig. 5, 6).

Fig. 5 One of silted habitats in Zaouïa, In-Salah

(photo crstra, 2013)

Fig. 6 Sanding of palm grove in In-Salah

(photo crstra, 2013)

3. DATA AND METHODS

The characterization of physiographic context of In-Salah region required the use of different

cartographic supports; topographical map of In-Salah sheet NG 31-2 at a scale of 1/250.000

(U.S. Army, 1953), geological map (Meyendorff and Follot, 1951) at a scale of 1/500.000,

Google earth, etc. Also, the climatic study was based on the series of rainfall, temperature,

wind and humidity, from the National Office of Meteorology for the periods 1973-2012 and

1998-2008.

The monitoring of the sanding evolution at In-Salah required processing of multitemporal

Landsat 5 and 7 images, corresponding to the scene center 194/041. The scenes 205-566 and

220-783 are acquired respectively on March 3. 1987 and April 13. 2005, and are remotely

sensed by the TM and ETM + sensors covering a period of 19 years. Their acquisition is

freely available via the web portal "Global Land Cover Facility". The choice of the Landsat

satellite was due to the medium spatial resolution of images (30 m), allowing analysis on a

regional scale. Selection of images depended mainly on their availability, as their quality and

spectral resolution (number of bands). These Landsat images were orthorectified,

georeferenced, radiometrically corrected and projected in WGS 84, according to the UTM

system zone 31, Clarke ellipsoid 1880.

Moreover, detection of the space changing in In-Salah had required the images processing

under the ENVI 4.7 software:

The choice of multispectral bands for both scenes was achieved by using channel 1 (blue),

3 (red) and 4 (near infrared) for a better visualization of the space elements.

The extraction of the study area had allowed obtaining two scenes of equal size (836×497

pixel) covering an area of 374 km².

Visualization of the colour compositions than, determination of the space elements.

Sampling on field assisted by Google earth, location of test plots on the study area and

analysis of their separability, by calculating the transformed divergence separability

(Richards, 1999) for each Landsat scene, this test gave significant results for scenes TM

Nouar BOULGHOBRA, Toufik HADRI and Mamar BOUHANA / USING LANDSAT IMAGINARY …5

(Table 1) and ETM +; the values were around 2.00 in the majority, meaning the good

independence of the obtained classes.

Classification of the images using the supervised mode, and the maximum likelihood

method.

Post classification and developing of static map of the space elements, by determining the

occupied area for each landscape unit corresponding to 1987 and 2005.

Extraction and superposing the "mobile dunes" vector layer for the two scenes, to obtain

the dynamic map of the mobile dunes evolution from 1987 to 2005.

Appreciation of the space elements dynamics by calculating the changing rate (FAO,

1996), expressed as:

𝐼𝑐 (%) = [1 − (𝐴1

𝐴2)

1n⁄

] × 100

Where A1 and A2 correspond to the surface area of a class at the beginning and the end

of the period, and n is the number of years of this period.

Table 1. Test of transformed divergence separability for the scene TM, In-Salah 1987

MD SE HBT ST BS MS PG AE

MD

SE 1,78

HBT 2,00 2,00

ST 1,95 1,83 2,00

BS 1,85 1,81 1,99 1,99

MS 2,00 2,00 1,97 2,00 2,00

PG 2,00 2,00 2,00 2,00 2,00 2,00

AE 2,00 1,99 2,00 2,00 2,00 2,00 2,00

MD: Mobile Dunes ; SE: Sand Erg ; HBT: Habitation ST: Sand topping ; BS: Bare soil ; MS: Moist soil; PG: Palme grove

AE: Agricultural Enhancement.

4. RESULTS AND DISCUSSION

The visual analysis of the bands trichromy (1-3-4) of each Landsat scenes was particularly

significant and represents the field reality. The land investigation has allowed the distinction

of the main units of land use in the study area, according to the different degrees of

reflectance. We distinguished palm groves, habitations, moist soil, bare soil, and sand forms

(erg, mobile dunes and sand toppings) that appear yellow. Maps resulting from the supervised

classification by the method of maximum likelihood, allowed the identification of space

components in 1987 and 2005 (Table 2).

Table 2. Comparison of the space components at In-Salah on 1987 and 2005

Year

Space component

1987 2005

Km² % Km² %

Palm grove 10 3 20 5

Habitations 8 2 14 4

Mobile dunes 23 6 26 7

Sand Erg 89 24 151 40

Sand topping 119 32 29 8

Bare soil 114 31 109 29

Agricultural enhancement 4 1 0 0

Moist soils 8 2 25 7

Total 374 100 374 100

6

Bare soils and bedrock outcrops are generally located in north, and correspond to the

Tademaït tray glacis. The field slopes promote the erosion actions more than accumulation.

Urban areas such as palm groves occupy the low parts of the plain, because of the availability

of underground water resources (albian aquifer), evermore accompanied by the moist soils

resulting from the palms irrigation and waste water.

Sands and mobile dunes from Erg Sidi Moussa occupy large areas in the region, they move

in the direction of the efficient winds and deposited on the eastern sides of all palm groves

and agglomerations especially In-Salah; wich is the most vulnerable agglomeration The

severity of sanding is less important in the other agglomerations, because of their offset

position from the low lands of the plain (Fig. 7a, 7b).

Fig. 7a Classified Landsat image, In-Salah 1987

Fig. 7b Classified Landsat image, In-Salah 2005

Nouar BOULGHOBRA, Toufik HADRI and Mamar BOUHANA / USING LANDSAT IMAGINARY …7

Also, the diachronic monitoring of the landscape elements shows that the region has

undergone significant changes, which results in increasing most of the space units (Table 3).

Table 3. Changing rate of some space components, In-Salah 1987 to 2005

Land use area (km²)

FAO changing rate (%) 1987 2005

Moist soil 7,7 25,3 6,1 (increasing)

Palm grove 9,7 19,5 3,6 (increasing)

Habitations 7,7 13,6 3,0 (increasing)

Mobile dunes 22,9 25,8 0,6 (increasing)

The areas occupied by moist soils are characterized by an exceptional changing rate (6.1 %),

this is mainly due to the expansion of date palm areas (3.6 %) being the principal crop for the

local population (edaphic, climatic and socio economic raisons). Therefore, this leads to more

moist surfaces resulting from drainage. Additionally, because of the demographic pressure,

built-up areas have increased from 7.7 km² in 1987 to 13.6 km² in 2005, representing a growth

rate of 3 %. The In-Salah agglomeration in particular has expanded in all directions.

Therefore, several new buildings were buried in a short time. In the same way, the evolution

of sandy areas is clearly visible; the surface of the mobile dunes increases from 22.9 km² in

1987 to 25.8 km² in 2005 (Fig. 8), either a growth rate of 0.6 %.

Fig. 8 Spatial evolution of mobile dunes in In-Salah, 1987-2005

This high increase is mainly due to the anterior climatic factors that favored the sand

dynamics, the decrease of the relative humidity (26.9 to 25.9 %) and the increase of the

temperature (25.9 to 26.5 °C) made the soils more dries and more able to mobilization. Also

the increasing of the maximum wind speed (6.4 to 9.4 m/s) has significantly accentuated the

sand transport capacity (Table 4).

8

Table 4. Explanatory elements of the mobile dunes expansion, In-Salah from 1987 to 2005

Multidate comparison parameters 1973 to 1987 1988 to 2005

Mobile dunes area (km²) 22,9 (1987) 25,8 (2005)

Relative humidity (%) 26,9 25,9

Temperatures (oC) 25,9 26,5

Maximum wind speed (m/s) 6,4 9,4

5. CONCLUSION

The risk of sanding in the Tidikelt plain is related to different physical and climatic factors

which converge to the acceleration of wind activities. However, the unfortunate installation

of urban, agricultural and industrial structures in a high risk area has greatly increased the

vulnerability.

The use of satellite imagery highlights a progressive and continuous dynamics of the sand

expanses in the Tidikelt, representing subsequently real and perceptible risk that affects all

aspects of life in In-Salah region. The substantial and permanent loss of crops, habitats and

infrastructures is in constant aggravation and reveals the low resilience of the population in

terms of risk. Several actions against land sanding were launched by the different institutions

and citizens which led to failure for different reasons; inappropriate choice of control

techniques, of land to be developed or even plantation, still the main and real handicaps to an

efficient and durable protection.

REFERENCES

Arrar, A. (1989) Current issues and trends in irrigation with special reference to developing countries,

in: “resource conservation and desertification control in the near east. Report of the international

training course, DSE, FAO, GTZ, UNESCWA, Germany and Kingdom of Jordan.

Bensaid, A. (2006) SIG et télédétection pour l’étude de l’ensablement dans une zone aride : le cas de

la wilaya de Naâma (Algérie). Thèse de doctorat en géographie, UJF et laboratoire espace

géographique et aménagement du territoire, université Es-Senia-Oran, 325p.

Costea, G. & Haidu, I. (2010) Detection of recent spatial changes regarding landuse in small basins

from the Apuseni Natural Park. Geographia Technica, No. 2, pp. 11 – 17. ISSN: 1842-

5135 http://technicalgeography.org/pdf/2_2010/gt_2_2010.pdf.

Costea, G., Serradj, A. & Haidu, I. (2012) Forest cartography using Landsat imagery, for studying

deforestation over three catchments from Apuseni mountains, Romania. In: Advances in Remote

Sensing, Finite Differences and Information Security. Proceedings of the 8th WSEAS

International Conference on Remote Sensing, pp. 109-114, ISBN: 978-1-61804-127-2.

http://www.wseas.us/e-library/conferences/2012/Prague/ISFINRE/ISFINRE-17.pdf.

Dymond, C. C., Mladenoff, D J. & Radeloff V C. (2002) Phenological differences in tasseled cap

indices improve deciduous forest classification. Remote sensing of environment (80), 460-472.

FAO. (1996) Forest resources assessment 1990. Survey of tropical forest cover and study of change

processes. FAO forestry paper, 130. Rome. ISBN, 92-5-103808-2.

FAO. (2010) Lutte contre l’ensablement, l’exemple de la Mauritanie. Étude FAO foret 158, 89p.

Kasbadji Merzouk, N. (1999) Carte des vents de l’Algérie - résultats préliminaires -, revue des

énergies renouvelables : valorisation (1999), 209-214.

Kerbani, B. (2013) Étude de l’ensablement par image satellite Landsat à In-Salah, mémoire de master

en géomorphologie appliquée. Université des sciences et de technologie Houari Boumediene,

Alger. 85p.

Nouar BOULGHOBRA, Toufik HADRI and Mamar BOUHANA / USING LANDSAT IMAGINARY …9

Mainguet, M. & Dumay, F. (2006) Combattre l’érosion éolienne: un volet de la lutte contre la

désertification. Les dossiers thématiques du CSFD. No 3. Avril 2006. CSFD/Agropolis,

Montpellier, France. 44p.

Office national des statistiques. (2008) Recensement général de la population et de l'habitat 2008.

Wilaya de Tamanrasset, Algérie. En ligne : www.ons.dz.

ONU. (1994) Convention internationale sur la lutte contre la désertification dans les pays gravement

touchés par la sécheresse et/ou la désertification, en particulier en Afrique. Texte final de la

convention. 65p.

Richrads, J. A., (1999) Remote sensing digital image analysis. Springer-Verlag, Berlin, 240p.

Thomas, D. S. G., (1995) Desertification: causes and processes, in: encyclopedia of environmental

biology, Vol. 1, ed. W. A. Nierenberg, San Diego, Academic Press, 463-473.

Zabeirou, T., Guéro, M. & Amadou, B. (2004) Étude sur la dynamique de l’ensablement dans le

département de Maïné-Soroa, cartographie des situations de l’ensablement. Rapport de projet

d’appui à la gestion des ressources naturelles PAGRN, 40p.